��<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> <head> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /> <meta name="viewport" content="width=800" /> <title>Subhodh Kotekal</title> <link rel="stylesheet" type="text/css" media="screen,print" href="./index.css" /> <!--[if IE]><link rel='stylesheet' type='text/css' media='screen,print' href='index_files/indexIE.css'/><![endif]--> <!-- <script type="text/javascript" src="Scripts/iWebSite.js"></script> <script type="text/javascript" src="Scripts/iWebImage.js"></script> --> <script type="text/javascript" src="index_files/index.js"></script> </head> <body style="background: #ffffff; margin: 0pt; " onload="onPageLoad();"> <div style="text-align: center; "> <div style="margin-bottom: 0px; margin-left: auto; margin-right: auto; margin-top: 0px; overflow: hidden; position: relative; word-wrap: break-word; background: #ffffff; text-align: left; width: 800px; " id="body_content"> <div style="float: left; margin-left: 0px; position: relative; width: 800px; z-index: 0; " id="nav_layer"> <div style="height: 0px; line-height: 0px; " class="bumper"> </div> <div style="clear: both; height: 0px; line-height: 0px; " class="spacer"> </div> </div> <div style="height: 400px; margin-left: 0px; position: relative; width: 800px; z-index: 10; " id="header_layer"> <div style="height: 0px; line-height: 0px; " class="bumper"> </div> <!-- Name and affiliation --> <div id="id1" style="height: 300px; left: 0px; position: absolute; top: 18px; width: 480px; z-index: 1; " class="style_SkipStroke"> <div class="text-content style_External_400_300" style="padding: 0px; "> <div class="style"> <p style="padding-top: 25pt; " class="paragraph_style">Subhodh Kotekal<br /></p> <p class="paragraph_style"><br /></p> <p class="paragraph_style_2"> <i> Starting from August 2025, I am a Norbert Wiener Fellow at the Massachusetts Institute of Technology (MIT).</i> <br /><br /> I obtained my Ph.D. from the Department of Statistics at the University of Chicago in June 2025. My advisor was <a href="https://www.stat.uchicago.edu/~chaogao/">Chao Gao</a>. Prior to graduate school, I received a B.S. in Mathematics and in Statistics from the University of Chicago. </p> <p class="paragraph_style_2"> <br /> Institute for Data, Systems, and Society<br /> 77 Massachusetts Avenue<br /> Cambridge, MA 02139-4307<br /><br /> skotekal@mit.edu<br /></p> </div> </div> </div> <!-- Picture --> <div style="height: 160px; width: 160px; height: 160px; left: 533px; position: absolute; top: 85px; width: 158px; z-index: 1; " class="tinyText style_SkipStroke_1 stroke_0"> <img src="skotekal.jpg" alt="" style="border: none; height: 335px; width: 267px; " /> </div> </div> <hr style="margin-top: 40px;" /><br /> <div style="margin-left: 0px; position: relative; width: 800px; z-index: 5; " id="body_layer"> <div style="height: 0px; line-height: 0px; " class="bumper"> </div> <div id="id3" style = "z-index: 1; " class="style_SkipStroke_2"> <div class="text-content graphic_textbox_layout_style_default_External_708_220" style="padding: 0px; "> <div class="graphic_textbox_layout_style_default"> <p class="paragraph_style_2"> I am broadly interested in score-based generative AI, particularly diffusion models, and score matching. I have also been interested in developing provably optimal statistical and machine learning methods for signal and anomaly detection, particularly in high-dimensional settings of extremely low signal-to-noise ratio where prediction is impossible. </p> <br /><br /><br /> <!-- ===================================================== --> <p class="paragraph_style_1"><span style="line-height: 25px; " class="style_3">Papers</span> <hr style="height:2px;border-width:0;color:gray;background-color:gray"></hr></p> <p class="paragraph_style_5"> <span class="style_4">Diffusion models are optimal for hypothesis testing</span><br /> S. Kotekal<br /> In submission (Available upon request)<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Sharp optimality of simple, plug-in estimation of the Fisher information of a smoothed density</span> [<a href="https://openreview.net/forum?id=R6aLcBGBXf">openreview</a>]<br /> S. Kotekal<br /> International Conference on Machine Learning (ICML), 2025, to appear<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">From optimal score matching to optimal sampling</span> [<a href="https://arxiv.org/abs/2409.07032">arXiv</a>]<br /> Z. Dou, S. Kotekal, Z. Xu, and H. Zhou<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Locally sharp goodness-of-fit testing in sup norm for high-dimensional counts</span> [<a href="https://arxiv.org/abs/2409.08871">arXiv</a>]<br /> S. Kotekal, J. Chhor, and C. Gao<br /> Bernoulli, to appear<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Variance estimation in compound decision theory under boundedness</span> [<a href="https://papers.nips.cc/paper_files/paper/2024/hash/a7667ee5d545a43d2f0fda98863c260e-Abstract-Conference.html">proc</a>] [<a href="https://openreview.net/forum?id=HvCppnDykt&nesting=2&sort=date-desc">openreview</a>]<br /> S. Kotekal<br /> Advances in Neural Information Processing Systems (NeurIPS), 2024<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Optimal estimation of the null distribution in large-scale inference</span> [<a href="https://arxiv.org/abs/2401.06350">arXiv</a>] [<a href="https://doi.org/10.1109/TIT.2025.3529457">doi</a>]<br /> S. Kotekal and C. Gao<br /> IEEE Transactions on Information Theory, 2025<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Sparsity meets correlation in Gaussian sequence model</span> [<a href="https://arxiv.org/abs/2312.09356">arXiv</a>] [<a href="http://dx.doi.org/10.1214/25-AOS2494">doi</a>]<br /> S. Kotekal and C. Gao<br /> Annals of Statistics, 2025<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Optimal heteroskedasticity testing in nonparametric regression</span> [<a href="https://arxiv.org/abs/2310.12424">arXiv</a>] [<a href="http://dx.doi.org/10.1214/24-AOS2467">doi</a>]<br /> S. Kotekal and S. Kundu<br /> Annals of Statistics, 2025<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Minimax signal detection in sparse additive models</span> [<a href="https://arxiv.org/abs/2304.09398">arXiv</a>] [<a href="https://doi.org/10.1109/TIT.2024.3473770">doi</a>]<br /> S. Kotekal and C. Gao<br /> IEEE Transactions on Information Theory, 2024<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Minimax rates for sparse signal detection under correlation</span> [<a href="https://arxiv.org/abs/2110.12966">arXiv</a>] [<a href="https://doi.org/10.1093/imaiai/iaad044">doi</a>]<br /> S. Kotekal and C. Gao<br /> Information and Inference, 2023<br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Statistical limits of sparse mixture detection</span> [<a href="https://arxiv.org/abs/2104.02507">arXiv</a>] [<a href="https://doi.org/10.1214/22-EJS2053">doi</a>]<br /> S. Kotekal<br /> Electronic Journal of Statistics, 2022 <br /> <br /> </p> <p class="paragraph_style_5"> <span class="style_4">Recurrent interactions can explain the variance in single trial responses</span> [<a href="https://doi.org/10.1371/journal.pcbi.1007591">doi</a>]<br /> S. Kotekal and J. N. MacLean<br /> PLOS Computational Biology, 2020 <br /> </p> <br /><br /><br /> </body> </html>